Thu, 03 Dec 2015 13:22:36 -0800 localrepo: add a second auditor without file system check
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 03 Dec 2015 13:22:36 -0800] rev 27233
localrepo: add a second auditor without file system check Auditors keeps a cache of audited paths. Therefore we cannot use the same auditor for working copy and history operation. We create a new one without file system check for this purposes.
Thu, 03 Dec 2015 10:40:19 -0800 pathauditor: add a way to skip file system check
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 03 Dec 2015 10:40:19 -0800] rev 27232
pathauditor: add a way to skip file system check We need to be able to skip it when looking at data within the history. Doing them in all cases leads to buggy behavior like issue4749.
Thu, 03 Dec 2015 12:22:48 -0800 pathauditor: move file system specific check in their own function
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 03 Dec 2015 12:22:48 -0800] rev 27231
pathauditor: move file system specific check in their own function This will make it easy to disable that part when not relevant (eg: auditing filename for operation in history)
Sat, 07 Nov 2015 16:31:04 +0900 contrib: disable SSLv3_method() to build old Python with recent libssl
Yuya Nishihara <yuya@tcha.org> [Sat, 07 Nov 2015 16:31:04 +0900] rev 27230
contrib: disable SSLv3_method() to build old Python with recent libssl Because OpenSSL is compiled without SSLv3 support on Debian sid, Python 2.6.9 can't be built without this hack. Python 2.7 is patched appropriately, but 2.6 isn't. http://bugs.python.org/issue22935
Thu, 03 Dec 2015 08:31:20 -0800 rebase: only clear rebase status after the rebase transaction has completed
Laurent Charignon <lcharignon@fb.com> [Thu, 03 Dec 2015 08:31:20 -0800] rev 27229
rebase: only clear rebase status after the rebase transaction has completed In 405320cd6198, I made the mistake of moving the step "clearing the status after a rebase" to inside the rebase transaction. This was wrong, since we don't want to clear the status (and the rebase state) if something went wrong during the transaction: if something goes wrong we want to keep the rebase state to be able to abort. It broke rebase with evolve + inhibit.
Thu, 03 Dec 2015 01:38:21 +0100 dirstate: don't write repo.currenttransaction to repo.dirstate if repo stable
Sietse Brouwer <sbbrouwer@gmail.com> [Thu, 03 Dec 2015 01:38:21 +0100] rev 27228
dirstate: don't write repo.currenttransaction to repo.dirstate if repo is None (issue4983) Some hooks, such as post-init and post-clone, do not get a repo parameter in their environment. If there is no repo, there is no repo.currenttransaction(); attempting to retrieve it anyway was causing crashes. Now currenttransaction is only retrieved and written if the repo is not None.
Wed, 02 Dec 2015 14:20:13 -0800 tests: test changegroup generation for filtered changesets (issue4982) stable
Gregory Szorc <gregory.szorc@gmail.com> [Wed, 02 Dec 2015 14:20:13 -0800] rev 27227
tests: test changegroup generation for filtered changesets (issue4982) The test demonstrates the buggy behavior from issue4982 where the changegroup contains changesets it shouldn't.
Wed, 02 Dec 2015 23:04:58 +0900 parsers: fix parse_dirstate to check len before unpacking header (issue4979) stable
Yuya Nishihara <yuya@tcha.org> [Wed, 02 Dec 2015 23:04:58 +0900] rev 27226
parsers: fix parse_dirstate to check len before unpacking header (issue4979)
(0) -10000 -3000 -1000 -300 -100 -30 -10 -8 +8 +10 +30 +100 +300 +1000 +3000 +10000 tip